Programmers learning Rust struggle to understand ownership types, Rust's core mechanism for ensuring memory safety without garbage collection. This paper describes our attempt to systematically design a pedagogy for ownership types.

1. We studied Rust developers' misconceptions of ownership to create the Ownership Inventory, a new instrument for measuring a person's knowledge of ownership. We found that Rust learners could not connect Rust's static and dynamic semantics, such as determining why an ill-typed program would (or would not) exhibit undefined behavior.
2. We created a conceptual model of Rust's semantics that explains borrow checking in terms of flow-sensitive permissions on paths into memory.
3. We implemented a Rust compiler plugin that visualizes programs under the model.
4. We integrated the permissions model and visualizations into a broader pedagogy of ownership by writing a new ownership chapter for The Rust Programming Language, a popular Rust textbook.
5. We evaluated an initial deployment of our pedagogy against the original version, using reader responses to the Ownership Inventory as a point of comparison. Thus far, the new pedagogy has improved learner scores on the Ownership Inventory by an average of 9% (N = 342, d = 0.56).